 |
 |
Manipulating Magnetic Field Shape
Name: Michael
Status: student
Age: N/A
Location: N/A
Country: N/A
Date: N/A
Question:
Is it possible to manipulate the magnetic lines of flux in such
a way as the lines extend into an oval, with the magnet towards one end of
the oval? If possible, please tell me how it might be done, or refer me to
some where with the information.
Replies:
First, I want to clarify if you mean field lines or flux lines.
Typically flux lines represent a magnetic field perpendicular to a
surface, represented as vectors on a surface -- since a vector is just
a magnitude and a direction, it cannot be an oval. The ovals you
commonly see around a magnet are field lines, not flux lines.
Are you saying you want the field lines to go around the object, but
not into it? Then, no it is not possible -- field lines emanate from an
object with a magnetic field. If you are just saying you want to design
an arbitrary field shape (in this case, an oval) around your magnet,
then it becomes an engineering problem of orienting and controlling
the magnetic moment of the various ferromagnetic domains in your
magnet. You just need a more highly magnetized area on one side than
the other. I am assuming you mean a ferromagnet -- although the answer
is largely similar if you mean other kinds of magnets, but the way you
magnetize it is different.
As for practically how you should do that, there are custom magnets
you can buy on the Internet, which might be the easiest way -- but if
you want to do it yourself, you will have a challenging task of
sintering magnetic powder then applying various levels of different
magnetic fields in order to 'magnetize' according to the field shape
you want.
Hope this helps,
Burr Zimmerman
Well, Michael, sounds like it might not be possible.
In empty space, magnetic field lines (and electric)
tend to arrange themselves into minimum-energy positions
which tend to have a roughly spherical intensity distribution.
Energy-density of either field goes as square of field strength in Tesla (or v/m)
Number of field lines are fixed by the ampere-turns or the trapped static electric charges
which caused the field to exist in the first place.
The equations which govern the preferred distribution derive from those two things.
To modify the preferred distribution requires something to modify the permeability
(or permitivity) of the space. Only matter does that; energy does not.
so making a magnetic field reach out farther in one direction than another
would require some magnetic matter on one side to help propagation
(like transformer iron or mu-metal)
or on the other side to hinder propagation (like superconductor).
Air-core copper-wire solenoid coils shape it too,
but only by adding "new" magnetic flux of their own.
The key thing is that matter is needed for any of that.
As an example you can imagine a bar magnet being gradually bent into a horseshoe magnet.
The strength distribution of a field around a bar magnet is pretty symmetric,
either spherical or ellipsoidal up to 2:1. I forget which.
Look up equations for a "magnetic dipole" field.
When it is bent, say, tips toward the right,
perhaps the field reaches out to the right a little more than to the left, but not by much.
Mostly it is like a dipole stretching across the empty space between the two right-pointing tips.
The permanent-magnet material has helped the flux-lines transit from tip1 to tip2
travelling a U-shaped path inside the magnet metal,
and matter can do that kind of thing.
No "remote influence on empty space" can do that,
with the possible exception of gross general-relativistic gravitational spatial distortions,
(such as proximity to black holes, or science-fiction warp-drives)
which are not really available for practical use in this century.
I suppose a dipole field around an object travelling near light speed would be compressed in the direction of travel,
by special relativity. Not too complicated.
But it would still be symmetric. Reach in opposite directions would be equal.
Jim Swenson
Click here to return to the Physics Archives
| |
Update: June 2012
|
|
